If you trade on Solana — manually, via a bot, or even just clicking swap buttons in a wallet — there's a non-trivial chance MEV bots are extracting value from your trades. Most people don't realise it because the extraction is silent: you submit a trade, it executes at a price that looks normal, and the difference between "what the AMM math should have given you" and "what you got" goes into a bot's wallet somewhere.
This piece covers what Solana MEV bots actually do, how to recognise the patterns when they're targeting you, and what changes when you route through Anti-MEV protection.
What MEV Bots Actually Are
MEV stands for Maximal Extractable Value — value that an actor with transaction-ordering power can capture beyond standard fees. On chains with public mempools (Ethereum, BSC), MEV bots watch the mempool, identify profitable extraction opportunities, and submit transactions that capture them.
Solana is structurally different. There's no traditional public mempool — transactions go directly to the next-scheduled validator. But MEV still exists, just in different shapes:
- Validator-side ordering MEV. The validator producing a block can reorder transactions in its slot. A bot working with (or paying) validators can effectively reorder.
- RPC-side observation MEV. RPC providers see transactions before forwarding to validators. A malicious or compromised RPC, or just one that exposes data carelessly, can leak transaction info to bots in time for them to react.
- Pre-confirmation visibility. Some transaction submission paths are observable to bots before the transaction lands. This is where most Solana MEV happens in practice.
The end result for traders is the same as on Ethereum: you submit a profitable swap, a bot front-runs you, you execute at the inflated price, the bot back-runs you, and the difference goes to the bot.
The Sandwich Attack: Step by Step
The dominant MEV pattern on Solana is the sandwich attack:
- You decide to swap 100 SOL into a memecoin.
- You submit the transaction. It's now visible to a sandwich bot before it lands.
- The sandwich bot front-runs you: buys the same memecoin first, pushing the price up.
- Your transaction executes at the inflated price. You get fewer tokens than the AMM math at submission would have suggested.
- The sandwich bot back-runs you: sells the position they just built, capturing the price impact your transaction created.
The bot's profit is the difference between the inflated price they sold at and the original price they bought at, minus their own fees. Your loss is the difference between the AMM-math expected output and what you actually received.
Volume on this is large. Anyone trading on Solana through a non-protected path is contributing to that volume.
How to Tell If You're Being Sandwiched
For traders running bots, the diagnostic is straightforward: compare the AMM-math expected output for each swap against the actual fill. If the gap is systematically larger than slippage tolerance would predict — especially on profitable swaps — you're being sandwiched.
Concrete signs:
- Profitable in dry-run / backtest, breakeven or worse in production
- Persistent slippage that exceeds the AMM curve calculation
- Worse fills on larger positions (sandwich bots scale with target size)
- Worse fills during congestion (more bot activity)
- "Successful" transactions whose actual outcomes don't match expectations
For manual traders, the diagnostic is harder because you don't usually have AMM-math expected outputs computed. Aggregators like bubblemaps and Solana-specific MEV explorers can sometimes flag specific transactions as sandwiched, but the most reliable signal is: route through an Anti-MEV path and see if your effective prices improve.
Solana MEV Bot Strategies in Detail
The category includes more than sandwich attacks:
Sandwich attacks. The dominant strategy. Front-run, let the target execute, back-run. Most economically meaningful MEV on Solana.
Just-in-time (JIT) liquidity. A bot adds liquidity to a concentrated AMM right before a large swap, captures fees, removes liquidity. Less harmful to the trader but technically a form of MEV.
Liquidation MEV. Watching lending protocols for unhealthy positions; liquidating them to capture liquidation fees. Affects borrowers but not traders directly.
Backrunning arbitrage. A trader's swap creates a price imbalance across pools; a bot arbitrages it back. This is technically MEV but usually doesn't directly harm the original trader.
NFT mint sniping. Bots that compete with humans for limited NFT mints. Same submission-stack mechanics, different domain.
For most traders, sandwich attacks are the only MEV they care about. The others are either someone else's problem or arguably beneficial to ecosystem health.
How Anti-MEV Protection Actually Works
There are several patterns. The most effective for traders:
Private RPC routing. The transaction never enters a public observation surface. From your client to a private endpoint to a validator, with no intermediate visibility. Bots can't front-run what they can't see.
This is what BoltTx does by default. There's no "MEV mode" toggle, no separate product, no opt-in. Every transaction routed through bolttx.io is delivered without pre-confirmation exposure.
Other patterns exist but have trade-offs:
MEV-blocker pattern. RPC routes through searchers who agree not to front-run, with refunds if MEV is extracted. Originated on Ethereum; less standardised on Solana. Adds coordination overhead.
Batch auctions. Orders collected over a window, matched together. Solana lacks a direct equivalent at scale.
Frontend-level Anti-MEV. Some trading frontends bundle their own protection. Only works when you trade exclusively through that frontend. Doesn't help bots or programmatic submitters.
For programmatic submission (bots, dApps, AI agents), private RPC routing is the right answer. The integration cost is low (one URL change), the protection is automatic, and there's no per-transaction decision to remember.
See our Anti-MEV deep dive for the full breakdown.
The Economic Impact
For active traders without protection, sandwich tax is meaningful. Specific numbers vary by:
- Position size (larger size = more sandwich-able)
- Pool depth (shallower = more sandwich-able)
- Token volatility (higher volatility = larger windows for extraction)
- Time of day (more bot activity at certain times)
Reasonable estimates: 5-30 bps per swap on heavily-traded pairs, more on memecoin pairs. For a bot doing thousands of swaps per day, this compounds quickly.
Closing this gap is the single largest economic improvement most active Solana traders can make. It costs nothing extra (Anti-MEV routing isn't a premium feature on the right RPC) and the ROI is immediate.
What to Do This Week
If you suspect or know you're losing money to MEV bots:
- Measure your sandwich tax. Compare AMM-math expected outputs to actual fills on recent swaps. Systematic deviation = sandwich exposure.
- Switch to an Anti-MEV RPC. Don't run any directional trading through a public-routing RPC. The protection is at the routing layer, not an add-on.
- Re-run the same trades on the new RPC for a week. Compare the deltas. If your fills are systematically tighter on the protected path, you've quantified your savings.
- For bot operators: monitor sandwich exposure as a metric. Add a per-swap diagnostic that flags trades where actual fill diverged significantly from AMM-math expected.
- For dApp operators: route user transactions through Anti-MEV by default. Don't make this an opt-in setting users have to find.
- Don't over-engineer this. The fix is one URL change. The complexity is the opposite of the value.
Try BoltTx for Anti-MEV Protection
BoltTx routes every transaction with Anti-MEV protection by default. No separate product, no opt-in, no extra fee:
import { Connection } from "@solana/web3.js";
const connection = new Connection(
"https://bolttx.io/?api-key=YOUR_API_KEY",
"processed"
);
Free tier signup. Run real volume against it for a week. Compare your effective fills against your current setup. If sandwich tax drops measurably, you have your answer.
FAQ
Are Solana MEV bots legal? The activity itself isn't currently illegal in most jurisdictions. The economic effect is value transfer from traders to bot operators. Some argue it's a market efficiency feature; others argue it's parasitic. Regulation may evolve.
Will Solana fix MEV at the protocol level? Some protocol-level proposals exist; none are imminent. The current Anti-MEV solutions are application-layer (RPC routing).
Does Anti-MEV protection slow my transactions? On the right architecture, no. The protection is in routing, not an extra confirmation step. BoltTx's Anti-MEV routing is the same path that delivers sub-second confirmation.
What if I'm just a manual trader, not running bots? Same protection applies. Use a wallet that routes through an Anti-MEV RPC, or trade through a frontend that does. The economic logic is the same; you're just submitting fewer transactions.
Is Anti-MEV the same as sandwich protection? Sandwich attacks are one form of MEV. Anti-MEV protection covers sandwiches, front-running, and back-running. The terms are often used interchangeably for traders.